Re: [PATCH 1/8] btrfs: zoned: introduce block_group context for submit_eb_page()

[Christoph Hellwig <hch@infradead.org>]

On Mon, Jul 24, 2023 at 01:18:30PM +0900, Naohiro Aota wrote:
> For metadata write out on the zoned mode, we call
> btrfs_check_meta_write_pointer() to check if an extent buffer to be written
> is aligned to the write pointer.
> 
> We lookup for a block group containing the extent buffer for every extent
> buffer, which take unnecessary effort as the writing extent buffers are
> mostly contiguous.
> 
> Introduce "bg_context" to cache the block group working on.

As someone who already found the eb_context naming and handling in the
existing code highly confusing, I wonder if we should first add a new
eb_write_context structure, which initially contains the wbc
and eb_context pointers, and which also would grow the bg.  This
should make argument passing a little simpler, but most importantly
removes all the dealing with the double pointers that need a lot
of checking and clearing.

> diff --git a/fs/btrfs/zoned.c b/fs/btrfs/zoned.c
> index 5e4285ae112c..58bd2de4026d 100644
> --- a/fs/btrfs/zoned.c
> +++ b/fs/btrfs/zoned.c
> @@ -1751,27 +1751,33 @@ bool btrfs_check_meta_write_pointer(struct btrfs_fs_info *fs_info,
>  				    struct extent_buffer *eb,
>  				    struct btrfs_block_group **cache_ret)
>  {
>
>
> -	struct btrfs_block_group *cache;
> -	bool ret = true;
> +	struct btrfs_block_group *cache = NULL;

.. and independent of the above comment, I also found the "cache" and
"cache_ret" namings here very highly confusing.  What speaks again
using a bg-based naming that makes it clear what we're dealing with?